← BackPT-BR

Install Fluents.

One command. 30 seconds. No account required.

macOS

Two options on Mac, same license key works on both. Pick by use case — comparison below.

Transcription only

Recommended

Native Swift build. Recording + structured ata after the call. No audio driver needed.

$ curl -fsSL https://fluents.app/install-mac.sh | sh
  • ✓ Native ScreenCaptureKit (no BlackHole)
  • ✓ Per-project memory injected into the ata
  • ✓ Session history + Markdown ata
  • ✓ Auto-update banner inside the app
  • ○ Live translation (coming soon)
  • ○ Live interview coach (coming soon)

macOS 14+ (Sonoma) · UI in PT-BR only · Apple Silicon or Intel

Full

Python build

Same engine as the Win/Linux build. Includes live features (translation, CV-driven coach) on top of the post-call ata.

$ curl -fsSL https://fluents.app/install.sh | sh
  • ✓ Live translation during the call
  • ✓ Live coach: draft answer pulled from your CV
  • ✓ Overlay invisible in screen share
  • ✓ Post-call ata
  • ! Requires BlackHole (installed automatically)

macOS 12+ · UI in PT/EN · Apple Silicon or Intel

Not sure which one?

  • → Interviewing in English this week or next: Full (you need the live coach)
  • → Want to document client meetings as ata for Obsidian/Notion: Transcription only (lighter, faster)
  • → Undecided: start with Transcription only. If you need the live features later, install the Full build on the same Mac — your license key works on both.

Heads-up on language: the Transcription-only build ships with PT-BR UI for now (transcription itself works in any language — only the interface text is fixed). The Full build ships in PT/EN.

Windows

Python build

Paste in PowerShell (run as admin):

$ Set-ExecutionPolicy Bypass -Scope Process -Force; iwr https://www.fluents.app/install.ps1 -UseBasicParsing | iex

Installs Python 3.11+ via winget, downloads Fluents, creates a Start Menu shortcut. You'll be prompted to install VB-Audio Virtual Cable for system audio capture.

Include https:// — PowerShell defaults to HTTP when you omit the protocol, and the redirect to HTTPS fails silently.

Linux

Cloud pipelineSame quality as Mac

Paste in bash:

$ curl -fsSL https://fluents.app/install-linux.sh | sh

Dedicated Linux installer with the new cloud pipeline pre-activated: Speech v2 chirp_2 (mic) + Speech v1 latest_long with diarization (system) + Vertex Gemini 2.5 Pro 2-pass for ata generation. Same quality as the native Mac app. System audio uses PulseAudio/PipeWire monitor source. Tested on Ubuntu, Debian, Fedora, Arch.

Live coach + translation (interview-prep features) are available in the legacy install via install.sh. X11 recommended for screen-share invisibility — Wayland blocks it by design.

Why two builds on Mac? The native Swift app uses ScreenCaptureKit (no virtual audio driver) and Apple's latest transcription stack. The Python build delivers the full feature set (live coach, live translation) and is being ported to Swift over the next weeks. Same license key works across both.

After installing

  1. Open Fluents.
  2. Complete onboarding (name, CV, languages) — about 2 minutes.
  3. If you bought Pro, go to Settings → License and paste your key.
  4. Press ⌘ I before your next Zoom / Meet / Teams call.

Can't paste the license key?

Make sure you're running version 0.2+ (any install from Apr 21 2026 onward). Older versions had a different license flow. If you still hit issues, email hi@fluents.app.

Back to homepage
Powered byAraraHQ